Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “HIGH EXPLOSIVES”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,729 records · Page 96

Nanoscale reshaping of resonant dielectric microstructures by light-driven explosions

Femtosecond-laser-assisted material restructuring employs extreme optical intensities to localize the ablation regions. To overcome the minimum feature size limit set by the wave nature of photons, there is a need for new approaches to tailored material processing at the nanoscale. Here, we report the formation of deeply-subwavelength features in silicon, enabled by localized laser-induced phase explosions in prefabricated silicon resonators. Using short trains of mid-infrared laser pulses, we demonstrate the controllable formation of high aspect ratio (>10:1) nanotrenches as narrow as ~ $λ / 80$. The trench geometry is shown to be scalable with wavelength, and controlled by multiple parameters of the laser pulse train, such as the intensity and polarization of each laser pulse and their total number. Particle-in-cell simulations reveal localized heating of silicon beyond its boiling point and suggest its subsequent phase explosion on the nanoscale commensurate with the experimental data. The observed femtosecond-laser assisted nanostructuring of engineered microstructures (FLANEM) expands the nanofabrication toolbox and opens exciting opportunities for high-throughput optical methods of nanoscale structuring of solid materials.

42 ENGINEERING↗

SUB-CHANDRASEKHAR MASS MODELS FOR SUPERNOVAE

For carbon-oxygen white dwarfs accreting hydrogen or helium at rates in the range {approx}(1-10) x 10{sup -8} M{sub sun} yr{sup -1}, a variety of explosive outcomes is possible well before the star reaches the Chandrasekhar mass. These outcomes are surveyed for a range of white dwarf masses (0.7-1.1 M{sub sun}), accretion rates ((1-7) x 10{sup -8} M{sub sun} yr{sup -1}), and initial white dwarf temperatures (0.01 and 1 L{sub sun}). The results are particularly sensitive to the convection that goes on during the last few minutes before the explosion. Unless this convection maintains a shallow temperature gradient and unless the density is sufficiently high, the accreted helium does not detonate. Below a critical helium ignition density, which we estimate to be (5-10) x 10{sup 5} g cm{sup -3}, either helium novae or helium deflagrations result. The hydrodynamics, nucleosynthesis, light curves, and spectra of a representative sample of detonating and deflagrating models are explored. Some can be quite faint indeed, powered at peak for a few days by the decay of {sup 48}Cr and {sup 48}V. Only the hottest, most massive white dwarfs considered with the smallest helium layers, show reasonable agreement with the light curves and spectra of common Type Ia supernovae (SNe Ia). For the other models, especially those involving lighter white dwarfs, the helium shell mass exceeds 0.05 M{sub sun} and the mass of the {sup 56}Ni that is synthesized exceeds 0.01 M{sub sun}. These explosions do not look like ordinary SNe Ia or any other frequently observed transient.

79 ASTRONOMY AND ASTROPHYSICS↗

Visualizing the Three-Dimensional Arrangement of Hydrogen Atoms in Organic Molecules by Coulomb Explosion Imaging

Structure-sensitive methods based on femtosecond light or electron pulses are now making it possible to measure how molecular structures change during light-induced processes. Despite significant progress, high-fidelity imaging of nuclear positions remains a challenge even for relatively small molecular systems and, notably, regarding the positions of hydrogen atoms. As demonstrated in recent work, X-ray-induced Coulomb explosion imaging (CEI) may overcome this obstacle, as its sensitivity does not depend on the mass of the imaged atoms. The photoinduced ring opening of the heterocyclic molecule 2(5 H )-thiophenone has attracted recent interest. Here, in this work, we show that CEI offers a powerful route to imaging the peripheral H atoms in this molecule and thus, more generally, to tracking detailed nuclear motions (e.g., isomerizations) in organic molecules on ultrafast time scales. Specifically, we record momentum-space Coulomb explosion images that report on the three-dimensional positioning of all nuclei within the molecule, for instance, distinguishing H atoms in C–H bonds that lie within or are directed out of the plane defined by the heavy atoms. The prospect of imaging peripheral H atoms to probe photochemical dynamics is explored by coupling ab initio molecular dynamics with classical Coulomb explosion simulations, thereby differentiating potential photoproduct isomers, including those whose structures primarily differ in the position of the hydrogens.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Advanced release technologies program

The objective of the ARTS program was to develop lighter and less expensive spacecraft ordnance and release systems that answer to the requirements of a wide variety of spacecraft applications. These improvements were to be evaluated at the spacecraft system level, as it was determined that there were substantial system-level costs associated with the present ordnance and release subsystems. New, better devices were to be developed, then flight qualified, then integrated into a flight experiment in order to prove the reliability required for their subsequent use on high-reliability spacecraft. The secondary goal of the program was to quantify the system-level benefits of these new subsystems based upon the development program results. Three non-explosive release mechanisms and one laser-diode-based ordnance system were qualified under the program. The release devices being developed were required to release high preloads because it is easier to scale down a release mechanism than to scale it up. The laser initiator developed was required to be a direct replacement for NASA Standard Initiators, since these are the most common initiator in use presently. The program began in October, 1991, with completion of the flight experiment scheduled for February, 1994. This paper provides an overview of the ARTS program, discusses the benefits of using the ARTS components, introduces the new components, compares them with conventional systems and each other, and provides recommendations on how best to implement them.

Purdy, Bill↗

NTT, SPITZER, AND CHANDRA SPECTROSCOPY OF SDSSJ095209.56+214313.3: THE MOST LUMINOUS CORONAL-LINE SUPERNOVA EVER OBSERVED, OR A STELLAR TIDAL DISRUPTION EVENT?

The galaxy SDSSJ095209.56+214313.3 (SDSSJ0952+2143 hereafter) showed remarkable emission-line and continuum properties and strong emission-line variability first reported in 2008 (Paper I). The spectral properties and low-energy variability are the consequence of a powerful high-energy flare which was itself not observed directly. Here we report follow-up optical, near-infrared (NIR), mid-infrared (MIR), and X-ray observations of SDSSJ0952+2143. We discuss outburst scenarios in terms of stellar tidal disruption by a supermassive black hole, peculiar variability of an active galactic nucleus (AGN), and a supernova (SN) explosion, and possible links between these scenarios and mechanisms. The optical spectrum of SDSSJ0952+2143 exhibits several peculiarities: an exceptionally high ratio of [Fe VII] transitions over [O III], a dramatic decrease by a factor of 10 of the highest-ionization coronal lines, a very unusual and variable Balmer line profile including a triple-peaked narrow component with two unresolved horns, and a large Balmer decrement. The MIR emission measured with the Spitzer IRS in the narrow 10-20 {mu}m band is extraordinarily luminous and amounts to L {sub 10-20{mu}}{sub m} = 3.5 x 10{sup 43} erg s{sup -1}. The IRS spectrum shows a bump around {approx}11 {mu}m and an increase toward longer wavelengths, reminiscent of silicate emission. The strong MIR excess over the NIR implies the dominance of relatively cold dust. The pre- and post-flare NIR host galaxy colors indicate a nonactive galaxy. The X-ray luminosity of L {sub x,0.1-10keV} = 10{sup 41} erg s{sup -1} measured with Chandra is below that typically observed in AGNs. Similarities of SDSSJ0952+2143 with some extreme SNe suggest the explosion of a SN of Type IIn. However, an extreme accretion event in a low-luminosity AGN or inactive galaxy, especially stellar tidal disruption, remain possibilities, which could potentially produce a very similar emission-line response. If indeed a SN, SDSSJ0952+2143 is one of the most distant X-ray- and MIR-detected SNe known so far, the most MIR luminous, and one of the most X-ray luminous. It is also by far the most luminous (>10{sup 40} erg s{sup -1}) in high-ionization coronal lines, exceeding previous SNe by at least a factor of 100.

99 GENERAL AND MISCELLANEOUS↗

The Young and Nearby Normal Type Ia Supernova 2018gv: UV-optical Observations and the Earliest Spectropolarimetry

The nondetection of companion stars in SN Ia progenitor systems lends support to the notion of double-degenerate systems and explosions triggered by the merging of two white dwarfs. This very asymmetric process should lead to a conspicuous polarimetric signature. By contrast, observations consistently find very low continuum polarization as the signatures from the explosion process largely dominate over the pre-explosion configuration within several days. Critical information about the interaction of the ejecta with a companion and any circumstellar matter is encoded in the early polarization spectra. In this study, we obtain spectropolarimetry of SN 2018gv with the ESO Very Large Telescope at -13.6 days relative to the B-band maximum light, or ~5 days after the estimated explosion—the earliest spectropolarimetric observations to date of any SN Ia. These early observations still show a low continuum polarization (lesssim0.2%) and moderate line polarization (0.30% ± 0.04% for the prominent Si ii λ6355 feature and 0.85% ± 0.04% for the high-velocity Ca component). The high degree of spherical symmetry implied by the low-line and continuum polarization at this early epoch is consistent with explosion models of delayed detonations and is inconsistent with the merger-induced explosion scenario. Finally, the dense UV and optical photometry and optical spectroscopy within the first ~100 days after the maximum light indicate that SN 2018gv is a normal SN Ia with similar spectrophotometric behavior to SN 2011fe.

79 ASTRONOMY AND ASTROPHYSICS↗

iPTF14yb: The First Discovery of a Gamma-Ray Burst Afterglow Independent of a High-Energy Trigger

We report here the discovery by the Intermediate Palomar Transient Factory (iPTF) of iPTF14yb, a luminous (Msub>r ≈ ₋27.8 mag), cosmological (redshift 1.9733), rapidly fading optical transient. We demonstrate, based on probabilistic arguments and a comparison with the broader population, that iPTF14yb is the optical afterglow of the long-duration gamma-ray burst GRB140226A. This marks the rst unambiguous discovery of a GRB afterglow prior to (and thus entirely independent of) an associated high-energy trigger. We estimate the rate of iPTF14yb-like sources (i.e., cosmologically dis- tant relativistic explosions) based on iPTF observations, inferring an all-sky value of $R_{rel}$ = 610yr -1 (68% con dence interval of 110{2000 yr -1 ). Our derived rate is consistent (within the large uncer- tainty) with the all-sky rate of on-axis GRBs derived by the Swift satellite. Finally, we brie y discuss the implications of the nondetection to date of bona de \orphan" afterglows (i.e., those lacking de- tectable high-energy emission) on GRB beaming and the degree of baryon loading in these relativistic jets.

79 ASTRONOMY AND ASTROPHYSICS↗

PROGENITOR-EXPLOSION CONNECTION AND REMNANT BIRTH MASSES FOR NEUTRINO-DRIVEN SUPERNOVAE OF IRON-CORE PROGENITORS

We perform hydrodynamic supernova (SN) simulations in spherical symmetry for over 100 single stars of solar metallicity to explore the progenitor-explosion and progenitor-remnant connections established by the neutrino-driven mechanism. We use an approximative treatment of neutrino transport and replace the high-density interior of the neutron star (NS) by an inner boundary condition based on an analytic proto-NS core-cooling model, whose free parameters are chosen such that explosion energy, nickel production, and energy release by the compact remnant of progenitors around 20 M{sub Sun} are compatible with SN 1987A. Thus, we are able to simulate the accretion phase, initiation of the explosion, subsequent neutrino-driven wind phase for 15-20 s, and the further evolution of the blast wave for hours to days until fallback is completed. Our results challenge long-standing paradigms. We find that remnant mass, launch time, and properties of the explosion depend strongly on the stellar structure and exhibit large variability even in narrow intervals of the progenitors' zero-age main-sequence mass. While all progenitors with masses below {approx}15 M{sub Sun} yield NSs, black hole (BH) as well as NS formation is possible for more massive stars, where partial loss of the hydrogen envelope leads to weak reverse shocks and weak fallback. Our NS baryonic masses of {approx}1.2-2.0 M{sub Sun} and BH masses >6 M{sub Sun} are compatible with a possible lack of low-mass BHs in the empirical distribution. Neutrino heating accounts for SN energies between some 10{sup 50} erg and {approx}2 Multiplication-Sign 10{sup 51} erg but can hardly explain more energetic explosions and nickel masses higher than 0.1-0.2 M{sub Sun }. These seem to require an alternative SN mechanism.

79 ASTRONOMY AND ASTROPHYSICS↗

High pressure-temperature polymorphism of 1,1-diamino-2,2-dinitroethylene

Here, 1,1-diamino-2,2-dinitroethylene (FOX-7) is a low sensitivity energetic material with performance comparable to commonly used secondary explosives such as RDX and HMX. At ambient pressure, FOX-7 exhibits complex polymorphism with at least three structurally distinct phases (α, β, and γ). In this study, we have investigated the high pressure-temperature stability of FOX-7 polymorphs using synchrotron mid-infrared (MIR) spectroscopy. At ambient pressure, our MIR spectra and corresponding differential scanning calorimetry (DSC) measurements confirmed the known α → β (~110 °C) and β → γ (~160 °C) structural phase transitions; as well as, indicated an additional transition γ → δ (~210 °C), with the δ phase being stable up to ~251 degree C prior to decomposition. In situ MIR spectra obtained during isobaric heating at 0.9 GPa, revealed a potential α → β transition that could occur as early as 180 degree C, while β → β+δ phase transition shifted to ~300 °C with suppression of γ phase. Decomposition was observed slightly above 325 °C at 0.9 GPa..

36 MATERIALS SCIENCE↗

Constraining absolute neutrino masses via detection of galactic supernova neutrinos at JUNO

A high-statistics measurement of the neutrinos from a galactic core-collapse supernova is extremely important for understanding the explosion mechanism, and studying the intrinsic properties of neutrinos themselves. In this paper, we explore the possibility to constrain the absolute scale of neutrino masses m{sub ν} via the detection of galactic supernova neutrinos at the Jiangmen Underground Neutrino Observatory (JUNO) with a 20 kiloton liquid-scintillator detector. In assumption of a nearly-degenerate neutrino mass spectrum and a normal mass ordering, the upper bound on the absolute neutrino mass is found to be m{sub ν}<(0.83±0.24) eV at the 95% confidence level for a typical galactic supernova at a distance of 10 kpc, where the mean value and standard deviation are shown to account for statistical fluctuations. For comparison, we find that the bound in the Super-Kamiokande experiment is m{sub ν}<(0.94±0.28) eV at the same confidence level. However, the upper bound will be relaxed when the model parameters characterizing the time structure of supernova neutrino fluxes are not exactly known, and when the neutrino mass ordering is inverted.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Signatures of current loop coalescence in solar flares

The nonlinear coalescence instability of current carrying solar loops can explain many of the characteristics of the solar flares such as their impulsive nature, heating and high energy particle acceleration, amplitude oscillations of electromagnetic emission as well as the characteristics of 2-D microwave images obtained during a solar flare. The physical characteristics of the explosive coalescence of currents are presented in detail through computer simulation and theory. Canonical characteristics of the explosive coalescence are: (1) a large amount of impulsive increase of kinetic energies of electrons and ions; (2) simultaneous heating and acceleration of electrons and ions in high and low energy spectra; (3) ensuing quasi-periodic amplitude oscillations in fields and particle quantities; and (4) the double peak (or triple peak) structure in these profiles, participate in the coalescence process, yielding varieties of phenomena.

Sakai, J.↗

Fallout melt debris and aerodynamically-shaped glasses in beach sands of Hiroshima Bay, Japan

An intricate association of millimeter-sized, aerodynamically-shaped debris, including glass spherules, glass filaments, and composite-fused melt particles was recovered from beach sands on the shores of the Motoujina Peninsula in Hiroshima Bay, Japan. Based on optical microscopy, this debris comprises six morphological groups ranging from clear glasses to rubber-like constituents. Scanning electron microscopy and synchrotron X-ray microdiffraction revealed dominant aluminum, silicon and calcium (Al-Si-Ca) elemental composition with some iron, mainly in glass, associated with precipitates of mullite and anorthite microcrystals, hematite dendrites and iron-chromium globules, indicative of original temperature conditions >1800 °C. Aerodynamically-shaped fallout debris, including glass spherules described in this study, are generally produced by single high-energy catastrophic events, such as an extraterrestrial body impacting Earth or a nuclear explosion. Our work interprets the large volumes of fallout debris generated under extreme temperature conditions as products of the Hiroshima August 6th, 1945 atomic bomb aerial detonation. The chemical composition of the melt debris provides clues to their origin, particularly with regard to city building materials. This study is the first published record and description of fallout resulting from the destruction of an urban environment by atomic bombing.

38 RADIATION CHEMISTRY, RADIOCHEMISTRY, AND NUCLEA↗

Deacetylation and Mechanical Refining (DMR) and Deacetylation and Dilute Acid (DDA) Pretreatment of Corn Stover, Switchgrass, and a 50/50 Corn Stover/Switchgrass Blend

The most economical feedstock available to future lignocellulosic biorefineries may be a feedstock blend. Corn stover, switchgrass, and a 50:50 blend of corn stover and switchgrass were subjected to three different dilute alkali deacetylation conditions in a 90 L reactor prior to either mechanical refining (DMR) in a disc refiner followed by a Szego mill or dilute acid pretreatment (DDA) at five different conditions in a 4 L batch steam explosion reactor. The forty-five DDA and nine DMR slurries were subjected to high solids enzymatic hydrolysis at an enzyme loading of 15 mg protein per gram cellulose using Novozymes Cellic CTec3 (DDA) or a 4:1 mix of Novozymes Cellic CTec3:HTec 3 (DMR). The highest glucose yield was achieved using DMR processed corn stover from the high severity deacetylation condition (92 °C and 100 kg NaOH/ODMT) at 95.9% ± 0.8%, while the greatest monomeric glucose yields for switchgrass were between 75% to 80%. The glucose yields from saccharification of the 50:50 corn stover/switchgrass blend was approximately the average of the two individual feedstocks for both DDA and DMR processes. Statistical analysis showed that for a given feedstock type dilute alkali deacetylation severity had a significant impact (p < 0.05) on enzymatic hydrolysis yields, and for a given deacetylation severity the feedstock type had a significant effect on glucose conversion with the addition of switchgrass resulting in lower glucose yields. A blending strategy to maximize enzymatic hydrolysis sugar yields while reducing carbohydrate solubilization during the alkaline treatment step was proposed that allows deacetylation of feedstocks with different recalcitrance at optimal conditions for that feedstock, then blending prior to enzymatic hydrolysis.

BIOMASS FUELS,INORGANIC, ORGANIC, PHYSICAL, AND AN↗

NUCLEOSYNTHESIS IN TWO-DIMENSIONAL DELAYED DETONATION MODELS OF TYPE Ia SUPERNOVA EXPLOSIONS

For the explosion mechanism of Type Ia supernovae (SNe Ia), different scenarios have been suggested. In these, the propagation of the burning front through the exploding white dwarf (WD) star proceeds in different modes, and consequently imprints of the explosion model on the nucleosynthetic yields can be expected. The nucleosynthetic characteristics of various explosion mechanisms are explored based on three two-dimensional explosion simulations representing extreme cases: a pure turbulent deflagration, a delayed detonation following an approximately spherical ignition of the initial deflagration, and a delayed detonation arising from a highly asymmetric deflagration ignition. Apart from this initial condition, the deflagration stage is treated in a parameter-free approach. The detonation is initiated when the turbulent burning enters the distributed burning regime. This occurs at densities around 10{sup 7} g cm{sup -3}-relatively low as compared to existing nucleosynthesis studies for one-dimensional spherically symmetric models. The burning in these multidimensional models is different from that in one-dimensional simulations as the detonation wave propagates both into unburned material in the high-density region near the center of a WD and into the low-density region near the surface. Thus, the resulting yield is a mixture of different explosive burning products, from carbon-burning products at low densities to complete silicon-burning products at the highest densities, as well as electron-capture products synthesized at the deflagration stage. Detailed calculations of the nucleosynthesis in all three models are presented. In contrast to the deflagration model, the delayed detonations produce a characteristic layered structure and the yields largely satisfy constraints from Galactic chemical evolution. In the asymmetric delayed detonation model, the region filled with electron capture species (e.g., {sup 58}Ni, {sup 54}Fe) is within a shell, showing a large off-set, above the bulk of {sup 56}Ni distribution, while species produced by the detonation are distributed more spherically.

79 ASTRONOMY AND ASTROPHYSICS↗

POSSIBLE DISINTEGRATING SHORT-PERIOD SUPER-MERCURY ORBITING KIC 12557548

We report on the discovery of stellar occultations, observed with Kepler, which recur periodically at 15.685 hr intervals, but which vary in depth from a maximum of 1.3% to a minimum that can be less than 0.2%. The star that is apparently being occulted is KIC 12557548, a V = 16 mag K dwarf with T{sub eff,s} {approx_equal} 4400 K. The out-of-occultation behavior shows no evidence for ellipsoidal light variations, indicating that the mass of the orbiting object is less than {approx}3 M{sub J} (for an orbital period of 15.7 hr). Because the eclipse depths are highly variable, they cannot be due solely to transits of a single planet with a fixed size. We discuss but dismiss a scenario involving a binary giant planet whose mutual orbit plane precesses, bringing one of the planets into and out of a grazing transit. This scenario seems ruled out by the dynamical instability that would result from such a configuration. We also briefly consider an eclipsing binary, possibly containing an accretion disk, that either orbits KIC 12557548 in a hierarchical triple configuration or is nearby on the sky, but we find such a scenario inadequate to reproduce the observations. The much more likely explanation-but one which still requires more quantitative development-involves macroscopic particles escaping the atmosphere of a slowly disintegrating planet not much larger than Mercury in size. The particles could take the form of micron-sized pyroxene or aluminum oxide dust grains. The planetary surface is hot enough to sublimate and create a high-Z atmosphere; this atmosphere may be loaded with dust via cloud condensation or explosive volcanism. Atmospheric gas escapes the planet via a Parker-type thermal wind, dragging dust grains with it. We infer a mass-loss rate from the observations of order 1 M{sub Circled-Plus} Gyr{sup -1}, with a dust-to-gas ratio possibly of order unity. For our fiducial 0.1 M{sub Circled-Plus} planet (twice the mass of Mercury), the evaporation timescale may be {approx}0.2 Gyr. Smaller mass planets are disfavored because they evaporate still more quickly, as are larger mass planets because they have surface gravities too strong to sustain outflows with the requisite mass-loss rates. The occultation profile evinces an ingress-egress asymmetry that could reflect a comet-like dust tail trailing the planet; we present simulations of such a tail.

79 ASTRONOMY AND ASTROPHYSICS↗

ULTRA-STRIPPED TYPE Ic SUPERNOVAE FROM CLOSE BINARY EVOLUTION

Recent discoveries of weak and fast optical transients raise the question of their origin. We investigate the minimum ejecta mass associated with core-collapse supernovae (SNe) of Type Ic. We show that mass transfer from a helium star to a compact companion can produce an ultra-stripped core which undergoes iron core collapse and leads to an extremely fast and faint SN Ic. In this Letter, a detailed example is presented in which the pre-SN stellar mass is barely above the Chandrasekhar limit, resulting in the ejection of only ∼0.05-0.20 M {sub ☉} of material and the formation of a low-mass neutron star (NS). We compute synthetic light curves of this case and demonstrate that SN 2005ek could be explained by our model. We estimate that the fraction of such ultra-stripped to all SNe could be as high as 10{sup –3}-10{sup –2}. Finally, we argue that the second explosion in some double NS systems (for example, the double pulsar PSR J0737–3039B) was likely associated with an ultra-stripped SN Ic.

79 ASTRONOMY AND ASTROPHYSICS↗

Tensorized Feature Spaces for Feature Explosion

In this paper 1 1 This manuscript has been authored by UT-Battelle, LLC under Contract No. DE-AC05-000R22725 with the U.S. Department of Energy. The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this manuscript, or allow others to do so, for United States Government purposes. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/downloads/doe-public-access-plan) This research used resources of the Oak Ridge Leadership Computing Facility, which is a DOE Office of Science User Facility supported under Contract DE-AC05-000R22725., we present a novel framework that uses tensor factorization to generate richer feature spaces for pixel classification in hyperspectral images. In particular, we assess the performance of different tensor rank decomposition methods as compared to the traditional kernel-based approaches for the hyperspectral image classification problem. We propose Orion, which takes as input a hyperspectral image tensor and a rank and outputs an enhanced feature space from the factor matrices of the decomposed tensor. Our method is a feature explosion technique that inherently maps low dimensional input space in $\mathbb{R}^{K}$ to high dimensional space in $\mathbb{R}^{R}$ , where $R\gg K$ , say in the order of 1000x, like a kernel. We show how the proposed method exploits the multi-linear structure of hyperspectral three dimensional tensor. We demonstrate the effectiveness of our method with experiments on three publicly available hyperspectral datasets with labeled pixels and compare their classification performance against traditional linear and non-linear supervised learning methods such as SVM with Linear, Polynomial, RBF kernels, and the Multi-Layer Perceptron model. Finally, we explore the relationship between the rank of the tensor decomposition and the classification accuracy using several hyperspectral datasets with ground truth.

Pasricha, Ravdeep Singh↗